Across 22 shared benchmarks, Kimi K3 scores higher on 22 and Qwen3.5 122B A10B on 0. The widest gap is ZeroBench, where Kimi K3 scores 0.4 against 0.1. Qwen3.5 122B A10B is the cheaper of the two on tracked API pricing ($0.40 against $3.00 per million input tokens).
| Benchmark | Kimi K3 | Qwen3.5 122B A10B |
|---|---|---|
| AA Agentic Index | 54.3 | 21.3 |
| AA Intelligence | 60 | 32.8 |
| AA-LCR | 82.7 | 70.3 |
| AA-Omniscience | 46 | -41.5 |
| Artificial Analysis Coding Index | 76.2 | 45.7 |
| baby_vision | 85.7 | 40.2 |
| browsecomp | 91.2 | 63.8 |
| coding_arena_elo | 1674 | 1358 |
| critpt | 23.4 | 0.9 |
| gdpval | 59.4 | 24.3 |
| GPQA Diamond | 93.5 | 86.6 |
| HLE | 56 | 47.5 |
| mathvision | 97.8 | 86.2 |
| MMMU-Pro | 83.4 | 76.9 |
| MMVU | 82.1 | 74.7 |
| OmniScience Accuracy | 47.6 | 24.4 |
| OmniScience Non-Hallucination | 49.1 | 12.9 |
| OSWorld-Verified | 84.8 | 58 |
| scicode | 58.7 | 42 |
| TauBench V3 - Banking | 46 | 15.3 |
| Terminal-Bench 2.1 | 88.3 | 47.6 |
| ZeroBench | 0.4 | 0.1 |
Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing. Quoted rates are the price-setter row we currently track for each model — its direct or vendor-official listing where one exists (direct, alibaba-official), otherwise the lowest tracked offer.